Mlpack: Implement Batch Normalisation Layer

[Manthan-R-Sheth]

I would like to begin the task of implementing deep learning modules with the implementation of Batch Normalisation Layer. This is due to the wide advantage of using BatchNorm for DL tasks. It can be used for FFN as well as Convolution Networks.

@zoq @lozhnikov @rcurtin, shall i go ahead with this task?

[zoq]

BatchNorm would be a great to have, there is an open PR https://github.com/mlpack/mlpack/pull/955 which I think is a good start, so if you like, clone the code and resolve the remaining issues.

[sourabhvarshney111]

@Manthan-R-Sheth I think we have completed our work for this. Can we close it now? What say?

[zoq]

Right, let's close this one.